Winfried H writes:
What could be the reason, that the Sony A7s has “just” 12MP?
Something to do with the 4K and video optimization?
DIGLLOYD: By using a sensor that is just over 4K video resolution (minimal cropping), the image can be read natively off the sensor with no need for resampling, and the sensor can be optimized for video use knowing that no resampling is needed, and quite possibly optimized in other ways also. A higher-res sensor would require more cropping (self defeating) or more complex resampling (more CPU power needed to do that also) and would pose other optimization and operational challenges.
I don’t know what is state of the art now, but one need only look at the mangled Live View (magnified) of the Nikon D800 to see that the every-3rd-line line-skipping approach makes an awful mess of things (resampling was probably too CPU intensive and still would lead to artifacts). While extracting 4K video out of the 24-megapixel Sony A7 or 36-megapixel Sony A7R presumably is technically feasible (somehow), the issue might be as simple as heat/power and resampling speed (CPU speed).
The 12-megapixel state of the art sensor resolution in the A7s dovetails with 4K resolution demands, but it also extends the brightness (darkness) range for video and stills into previously problematic territory. I see it as part of a natural trend to build cameras that fit into certain usage needs better than jack-of-all-trades solutions.